Azure online endpoint is Scaling taking long time

### Operating System

Linux

### Version Information

Using the latest version of ML online endpoint

### Steps to reproduce

I have a project to run an AI model using an online endpoint as a backend service, the endpoint is configured (manually set in the portal) to be auto-scale based on the number of requests.


### Expected behavior

Expect the endpoint will scale up between 1-2 minutes like other services such as virtual machine scaleset, etc...

### Actual behavior

With endpoint, scaling takes a long time, about 12-18 minutes.

### Addition information

Do you have suggestions for speeding up the scaling time?

The issue provides no repository file, test, or implementation entry point. Start by reviewing the Azure Machine Learning online endpoint autoscale configuration in the portal and the reported 12–18 minute behavior; done would require a documented, reproducible cause or a confirmed way to reduce the scale-up time.

Written by the indexing model from the issue text.